feat(feedback): add /gsd-feedback with shared issue-filing pipeline #2371
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,55 @@ | ||
| --- | ||
| type: prompt | ||
| name: gsd:feedback | ||
| description: File GSD bugs, feature requests, or questions with diagnostics auto-attached | ||
| argument-hint: "[freeform description]" | ||
| allowed-tools: | ||
| - Read | ||
| - Bash | ||
| - AskUserQuestion | ||
| --- | ||
|
|
||
| <objective> | ||
| Collect user feedback from inside a GSD session, attach actionable diagnostics automatically, | ||
| and file a GitHub issue against `gsd-build/get-shit-done`. | ||
|
|
||
| Preferred path: create the issue with `gh`. | ||
| Fallback path: open a prefilled GitHub issue URL. | ||
| Final fallback: always return the full markdown body for manual paste. | ||
| </objective> | ||
|
|
||
| <execution_context> | ||
| @~/.claude/get-shit-done/workflows/feedback.md | ||
| </execution_context> | ||
|
|
||
| <context> | ||
| **Issue types:** | ||
| - `bug` | ||
| - `feature` | ||
| - `question` | ||
|
|
||
| **Repo target:** | ||
| - `gsd-build/get-shit-done` | ||
|
|
||
| **Input:** | ||
| - `$ARGUMENTS` may contain a freeform seed for the title and description | ||
| </context> | ||
|
|
||
| <process> | ||
| Read and execute the feedback workflow from @~/.claude/get-shit-done/workflows/feedback.md end-to-end. | ||
| </process> | ||
|
|
||
| <success_criteria> | ||
| - User intent captured as `bug`, `feature`, or `question` | ||
| - Diagnostics attached from current runtime and project state | ||
| - `gh issue create` attempted first when available | ||
| - Prefilled GitHub issue URL generated as fallback | ||
| - Full markdown body returned for manual filing even if submission fails | ||
| </success_criteria> | ||
|
|
||
| <critical_rules> | ||
| - Prefer machine-readable diagnostics from GSD helpers over ad hoc parsing | ||
| - Do not discard `$ARGUMENTS`; use it as a seed whenever possible | ||
| - Always return enough information for manual filing | ||
| - Never block on browser launch failure | ||
| </critical_rules> |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,162 @@ | ||
| <purpose> | ||
| Collect structured feedback, attach diagnostics from the current GSD session, and file a GitHub issue against `gsd-build/get-shit-done`. Issue submission is delegated to the shared `file-issue.md` workflow so that `/gsd-feedback` and `/gsd-forensics` use one canonical filing path. | ||
| </purpose> | ||
|
|
||
| <required_reading> | ||
| Read all files referenced by the invoking prompt's execution_context before starting. | ||
| </required_reading> | ||
|
|
||
| <process> | ||
|
|
||
| <step name="collect_feedback"> | ||
| Start from `$ARGUMENTS` if present. | ||
|
|
||
| If type is not already clear, ask the user to choose one: | ||
| - `bug` | ||
| - `feature` | ||
| - `question` | ||
|
|
||
| Then collect: | ||
| - `ISSUE_TITLE` | ||
| - `ISSUE_DESCRIPTION` | ||
|
|
||
| If `$ARGUMENTS` already contains enough detail, refine it instead of re-asking from scratch. | ||
|
|
||
| Keep the intake short. The goal is to get a fileable issue, not run a long interview. | ||
|
|
||
| ### Optional forensics enrichment (bug type only) | ||
|
|
||
| If `ISSUE_TYPE` is `bug`, ask the user: | ||
|
|
||
| > "Want me to run a quick forensics investigation first? This checks git history, planning state, and artifacts for anomalies that might explain the bug. [Y/n]" | ||
|
|
||
| If yes: | ||
| 1. Execute Steps 2–4 from `@~/.claude/get-shit-done/workflows/forensics.md` (Gather Evidence → Detect Anomalies → Generate Report). Skip Steps 5–8 (presentation, interactive investigation, issue creation, STATE update — those are handled here). | ||
| 2. Read the generated report from `.planning/forensics/report-{timestamp}.md`. | ||
| 3. Set `INVESTIGATION_FINDINGS` to the redacted report contents. | ||
|
|
||
| If no (or if `ISSUE_TYPE` is `feature` or `question`): | ||
| - Set `INVESTIGATION_FINDINGS` to empty string. | ||
| </step> | ||
|
|
||
| <step name="gather_diagnostics"> | ||
| Gather diagnostics from existing GSD helpers first. | ||
|
|
||
| **1. Version and runtime context** | ||
| Derive the invoking config dir from the execution context path when possible and prefer its sibling `get-shit-done/VERSION` file: | ||
|
|
||
| ```bash | ||
| PREFERRED_CONFIG_DIR="" | ||
| case "$PWD" in | ||
| *"/get-shit-done/workflows"*) | ||
| PREFERRED_CONFIG_DIR="${PWD%/get-shit-done/workflows*}" | ||
| ;; | ||
| esac | ||
|
|
||
| GSD_VERSION="" | ||
| for candidate in \ | ||
| "$PREFERRED_CONFIG_DIR/get-shit-done/VERSION" \ | ||
| "$HOME/.claude/get-shit-done/VERSION" \ | ||
| "$HOME/.codex/get-shit-done/VERSION" \ | ||
| ".claude/get-shit-done/VERSION" \ | ||
| ".codex/get-shit-done/VERSION" | ||
| do | ||
|
coderabbitai[bot] marked this conversation as resolved.
|
||
| if [ -n "$candidate" ] && [ -f "$candidate" ]; then | ||
| GSD_VERSION="$(cat "$candidate")" | ||
| break | ||
| fi | ||
| done | ||
|
|
||
| if [ -z "$GSD_VERSION" ] && [ -f package.json ]; then | ||
| GSD_VERSION=$(node -e "const fs=require('fs'); const pkg=JSON.parse(fs.readFileSync('package.json','utf8')); console.log(pkg.version || '')" 2>/dev/null) | ||
| fi | ||
|
|
||
| NODE_VERSION=$(node -v 2>/dev/null || echo "unknown") | ||
| OS_NAME=$(uname -s 2>/dev/null || echo "$OS") | ||
| ARCH_NAME=$(uname -m 2>/dev/null || node -p "process.arch" 2>/dev/null || echo "unknown") | ||
| CURRENT_DIR=$(pwd) | ||
| ACTIVE_WORKSTREAM=${GSD_WORKSTREAM:-none} | ||
| ``` | ||
|
|
||
| **2. State and phase diagnostics** | ||
| ```bash | ||
| STATE_JSON=$(gsd-sdk query state.json 2>/dev/null || echo "{}") | ||
| STATE_SNAPSHOT=$(gsd-sdk query state-snapshot 2>/dev/null || echo "{}") | ||
|
|
||
| CURRENT_PHASE=$(printf '%s' "$STATE_SNAPSHOT" | jq -r '.current_phase // "none"' 2>/dev/null || echo "none") | ||
| CURRENT_PHASE_NAME=$(printf '%s' "$STATE_SNAPSHOT" | jq -r '.current_phase_name // "none"' 2>/dev/null || echo "none") | ||
| STATUS=$(printf '%s' "$STATE_SNAPSHOT" | jq -r '.status // "unknown"' 2>/dev/null || echo "unknown") | ||
| LAST_ACTIVITY=$(printf '%s' "$STATE_SNAPSHOT" | jq -r '.last_activity // "unknown"' 2>/dev/null || echo "unknown") | ||
| LAST_ACTIVITY_DESC=$(printf '%s' "$STATE_SNAPSHOT" | jq -r '.last_activity_desc // "unknown"' 2>/dev/null || echo "unknown") | ||
| RESUME_FILE=$(printf '%s' "$STATE_SNAPSHOT" | jq -r '.session.resume_file // "none"' 2>/dev/null || echo "none") | ||
| ``` | ||
|
|
||
| **3. Small config excerpt** | ||
| Prefer direct GSD queries: | ||
|
|
||
| ```bash | ||
| MODEL_PROFILE=$(gsd-sdk query config-get model_profile 2>/dev/null || echo "unset") | ||
| BASE_BRANCH=$(gsd-sdk query config-get git.base_branch 2>/dev/null || echo "unset") | ||
| USE_WORKTREES=$(gsd-sdk query config-get workflow.use_worktrees 2>/dev/null || echo "unset") | ||
| COMMIT_DOCS=$(gsd-sdk query config-get commit_docs 2>/dev/null || echo "unset") | ||
| ``` | ||
|
|
||
| If `.planning/config.json` exists, include only a short relevant snippet rather than dumping the whole file. | ||
| </step> | ||
|
|
||
| <step name="render_diagnostics"> | ||
| Render the gathered diagnostics into `DIAGNOSTICS_MARKDOWN`: | ||
|
|
||
| ````markdown | ||
| - GSD version: {GSD_VERSION} | ||
| - Node: {NODE_VERSION} | ||
| - OS / arch: {OS_NAME} / {ARCH_NAME} | ||
| - Working directory: {CURRENT_DIR} | ||
| - Workstream: {ACTIVE_WORKSTREAM} | ||
| - Current phase: {CURRENT_PHASE} - {CURRENT_PHASE_NAME} | ||
| - Status: {STATUS} | ||
| - Last activity: {LAST_ACTIVITY} | ||
| - Last activity detail: {LAST_ACTIVITY_DESC} | ||
| - Resume file: {RESUME_FILE} | ||
| - Config excerpt: | ||
| - model_profile: {MODEL_PROFILE} | ||
| - git.base_branch: {BASE_BRANCH} | ||
| - workflow.use_worktrees: {USE_WORKTREES} | ||
| - commit_docs: {COMMIT_DOCS} | ||
|
|
||
| ### Raw State JSON | ||
|
|
||
| ```json | ||
| {STATE_JSON} | ||
| ``` | ||
|
|
||
| ### Raw State Snapshot | ||
|
|
||
| ```json | ||
| {STATE_SNAPSHOT} | ||
| ``` | ||
|
ElliotDrel marked this conversation as resolved.
Outdated
Comment on lines
+137
to
+141
There was a problem hiding this comment. Implement the redaction logic for Line 140 uses 🔧 Suggested implementation before line 117# Redact sensitive fields from state snapshot
STATE_SNAPSHOT_REDACTED=$(printf '%s' "$STATE_SNAPSHOT" | jq '
# Redact absolute paths - replace home directory
walk(if type == "string" then gsub(env.HOME; "~") else . end)
# Remove potentially sensitive keys
| del(.session.auth_token, .session.api_keys, .secrets)
# Truncate to reasonable size (keep first 50 keys/2KB)
' 2>/dev/null | head -c 2048)
# Add ellipsis if truncated
if [ ${`#STATE_SNAPSHOT_REDACTED`} -eq 2048 ]; then
STATE_SNAPSHOT_REDACTED="${STATE_SNAPSHOT_REDACTED}... [truncated]"
fiAdd this to the gather_diagnostics or render_diagnostics step. 🤖 Prompt for AI Agents |
||
| ```` | ||
| </step> | ||
|
|
||
| <step name="file_issue"> | ||
| Set the variables required by the shared issue-filing workflow, then delegate: | ||
|
|
||
| - `ISSUE_TYPE` — from step 1 | ||
| - `ISSUE_TITLE` — from step 1 | ||
| - `ISSUE_DESCRIPTION` — from step 1 | ||
| - `DIAGNOSTICS_MARKDOWN` — from step 3 | ||
| - `INVESTIGATION_FINDINGS` — from step 1 forensics enrichment (empty if skipped) | ||
| - `REPO` — `gsd-build/get-shit-done` | ||
|
|
||
| Execute `@~/.claude/get-shit-done/workflows/file-issue.md` end-to-end. It handles rendering the final issue body, `gh issue create`, URL fallback, and raw markdown fallback. | ||
| </step> | ||
|
|
||
| </process> | ||
|
|
||
| <success_criteria> | ||
| - Intake completes with type, title, and description | ||
| - Diagnostics come from `gsd-sdk query` helpers where available | ||
| - Bug-type issues offer optional forensics enrichment via `/gsd-forensics` steps 2–4 | ||
| - Issue filing delegates to `@~/.claude/get-shit-done/workflows/file-issue.md` | ||
| - `DIAGNOSTICS_MARKDOWN` and `INVESTIGATION_FINDINGS` are set before delegation | ||
| </success_criteria> | ||
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Uh oh!
There was an error while loading. Please reload this page.